How to Win
Box three of the Strategy Choice Cascade, and, in Martin’s own words, the hardest question in strategy, the most important, and the one most commonly faked.
How to Win is a theory of advantage: a specific, testable argument for why the chosen customer chooses you over every alternative, on your chosen field. The pairing matters, Where to Play and How to Win are one decision, because advantage only exists somewhere in particular. Nobody wins in general.
Why it gets faked
It gets faked because a list of sensible improvements feels like an answer. Faster, cheaper, higher quality, better service: every item defensible, every item something your board will nod at. But every competitor is pursuing the same list, which is why the list produces parity at rising cost rather than advantage. That path has a name in the library, Operational Excellence Is Not Strategy, and it is the most expensive way a company can avoid making a choice.
Three diagnostics
Run a claimed How to Win through three tests. The Opposite Test: could a rival sanely choose the reverse? If not, you have described an imperative, not an advantage. The Can't-Won't Test: is the advantage protected because rivals cannot copy it, or because they will not, since copying would break their own model? Both are real protection; knowing which you hold tells you how much time you have. And the logo test: remove your name from the strategy document, could a customer tell whose How to Win it is? If it could belong to anyone in the industry, it belongs to no one.
Where it has to cash out
A real theory of advantage ends in compelling customer action: customers doing something different, choosing, paying, staying, because of the advantage. That is Martin's deepest law, the customer is the only judge, and it disqualifies every How to Win whose evidence is internal enthusiasm. The advantage is sustained by must-have capabilities built specifically for it: not generic excellence, but the particular things you must be great at for this theory to hold.
How this connects to OKRs
This box is where our OKR methodology plugs into strategy most directly. Strategic cross-functional OKRs must trace to the How to Win, and Key Results are the evidence the theory is working: customer behaviour moving in the direction the advantage predicts. An OKR with no How to Win behind it is a goal, not a bet, and a quarter spent achieving it teaches you nothing about your strategy.
Our opinion: the Key Result is the receipt on a theory of advantage
This is the seam we care about most, so it is worth stating as a rule rather than a hint. A How to Win is a claim about why customers will choose you; a Key Result is the evidence the claim is holding. If your theory of advantage is lower switching cost, the KR is renewal rate moving from, say, 84% to 91%, not "ship the retention programme". The behaviour is the receipt; the programme is only the attempt. Our position, falsifiable on any dashboard: a portfolio whose Key Results are all internal milestones has no way of telling you whether its strategy is true, only whether its people were busy. That is the difference between a task-based Key Result and a real one.
One line to keep: A Key Result is the receipt on a theory of advantage: customer behaviour moving the way the advantage predicted.
Our synthesis of Roger Martin’s published work, sources credited. Read the originals: they’re excellent.
What makes a good How to Win Key Result? +
Customer behaviour moving the way your advantage predicts: renewal rate, switching, willingness to pay, referral. Not "launch the programme". The behaviour is the evidence the theory is true; the launch is only the attempt.
How can I tell a real How to Win from a faked one? +
Run three tests: could a rival sanely choose the opposite; is it protected because rivals can't or won't copy it; and if you removed your logo, could a customer still tell it was yours? A list of "faster, cheaper, better" fails all three.
Reading about method is not the same as running it. We install this system and build the capability that stays.
Turn strategy into executed strategy →